Here’s the Randomized Control Trial that I really want to see.
Take a lot of Type 2 Diabetics with BMIs in the obese range. Split them into two groups who are pair matched. Start one of the groups on Low Carb / High Fat diet and leave the control group on their customary Standard American Diet (SAD). Treat all of them with the standard of care as it is at the time. Track them for 40 years and look at the outcomes. Don’t just track some of the benchmarks like LDL cholesterol. Track all of their results including all-cause mortality.
It won’t happen for too many reasons. And it doesn’t take a belief in conspiracy theories to figure out why. Perhaps the biggest reason is nobody makes money with Low Carb/High Fat and a study with sufficient statistical power would be very expensive.
In the meanwhile, we are all n=1. And none of us have 40 years. And no point in looking for the RCT above since it’s never going to happen. It would have had to start before anyone knew the right questions to ask.
I’ve done somewhere around 50 grams of carbohydrates a day (30 grams when subtracting out fiber) for the past year. And my blood sugar control has been great. I’ve wondered how low someone has to go (or stay) in order to control Type 2 Diabetes. Certainly, the weight loss I’ve had (120 lbs) is a part of the solution. Being at a low body fat percentage now (7.5% per BodPod) has to help as well. My weight has been stable for 6 months now as well which means I’ve not lost or gained any weight – I was at 164.7 lbs when I took the BodPod test and today I weighed 163.7 lbs (close enough). My coffee consumption (which helps in weight loss for sure) is higher than ever before but I’m trying to keep the caffeine down by mixing in mostly decaf coffee.
So all of this begs the question of how many grams of carbs I could tolerate. Now, I am not going to be testing this anytime soon. I find the advantages of being low carb just way too easy (see above for the results). I did find a study that might provide at least a partial answer to the question of how many grams of carbs can keep someone in remission from Diabetes.
Now, I wouldn’t even call this diet at 30% of calories from carbohydrates a “Low Carbohydrate” by any definition that I would recognize but it has interesting results. One of the things that was interesting is that the study was done on severe diabetics (HbA1c levels of 9.0% or above). This is not a group of new diabetics nor were the participants young. They were a pretty good representative of Type 2 Diabetics with poor blood sugar control. The participants:
were instructed to follow a low-carbohydrate diet (1852 kcal; %CHO:fat:protein = 30:44:20) for 6 months in an outpatient clinic and were followed to assess their HbA1c levels, body mass index and doses of antidiabetic drugs.
The results were really good. Many of the participants got off their medications and:
HbA1c levels decreased sharply from a baseline of 10.9 ± 1.6% to 7.8 ± 1.5% at 3 months and to 7.4 ± 1.4% at 6 months.
These are similar to the results I got with the Low Carbohydrate diet when I got to an HbA1c level of 6.4. They are not as good as the results I got over the last couple of years with even lower levels of carbohydrates plus intermittent fasting.
In spite of being on a fairly low calorie diet (1852 kcal) they didn’t lose much weight. This group also didn’t seem to be all that obese since their BMI was around 24 (top end of “normal” weight).
Body mass index decreased slightly from baseline (23.8 ± 3.3) to 6months (23.5 ± 3.4).
So, if you are “normal” weight and diagnosed as diabetic then dropping from the Standard American Diet (SAD) 50% of calories from carbohydrates to 30% might give as good of control as exogenous Insulin without the long term increase in insulin resistance that comes along with Insulin therapy.
There was one telling outlier in the data.
One female patient had an increased physical activity level during the study period in spite of our instructions. However, her increase in physical activity was no more than one hour of walking per day, four days a week. She had implemented an 11%-carbohydrate diet without any anti-diabetic drug, and her HbA1c level decreased from 14.4% at baseline to 6.1% after 3 months and had been maintained at 5.5% after 6 months.
The paper presents the following lines of evidence for the claim,
Here we propose that glucagon excess, rather than insulin deficiency, is the sine qua non of diabetes. We base this on the following evidence:
(a) glucagon increases hepatic glucose and ketone production, catabolic features present in insulin deficiency;
(b) hyperglucagonemia is present in every form of poorly controlled diabetes;
(c) the glucagon suppressors leptin and somatostatin suppress all catabolic manifestations of diabetes during total insulin deficiency;
(d) total β cell destruction in glucagon receptor-null mice does not cause diabetes; and (e) perfusion of normal pancreas with anti-insulin serum causes marked hyperglucagonemia.
The insight that this may not be as much an insulin issue as a glucagon issue is a powerful one which may have application with medications to control Type 2 Diabetes. If giving exogenous insulin produces problems with Insulin Resistance, giving a medication which causes the body to produce less glucagon may have an opposite effect. It may be possible to develop a medication which downregulates glucagon indefinitely.
Here is a nice paper from 2009 on mice fed an ad libitum ketogenic diet (Kennedy AR, Pissios P, Otu H, Roberson R, Xue B, Asakura K, Furukawa N, Marino FE, Liu FF, Kahn BB, Libermann TA, Maratos-Flier E. A high-fat, ketogenic diet induces a unique metabolic state in mice. Am J Physiol Endocrinol Metab. 2007 Jun;292(6):E1724-39. Epub 2007 Feb 13).
The study looked at:
C57BL/6 mice animals were fed one of four diets:
2) a commonly used obesogenic high-fat, high-sucrose diet (HF);
3) 66% caloric restriction (CR); and
4) control chow (C).
Calories were the same but weight was lower on the ketogenic diet.
Mice on KD ate the same calories as mice on C and HF, but weight dropped and stabilized at 85% initial weight, similar to CR.
In fact, they moved mice from the High Fat High Carb diet to the Ketogenic diet and had the following:
Animals made obese on HF and transitioned to KD lost all excess body weight, improved glucose tolerance, and increased energy expenditure.
Even more along my own area of interest:
KD fed mice had a unique metabolic and physiological profile, exhibiting increased energy expenditure and very low respiratory quotient
The macronutrient composition of the diets was interesting:
Note this was not a high protein KD. I.e., The dietary advantage wasn’t protein. The percentage of calories from protein was the lowest on the KD – by far. This is a much higher level of fat than most people will tolerate and the protein level is pretty low.
Most telling was the body composition changes (Table 5).
The Chow fed mice were a bit over 10% heavier but at a lower % of Body Fat (13.5%) vs the Ketogenic fed mice. This can be attributed to the much lower protein consumption of the KD.
A contrasting study (Protein Leverage Hypothesis Counterpoint) showed an inflection point around 70% for fat where additional fat did not result in additional weight. In my opinion (study needed) – substituting protein for some of the fat should not be an issue.
The study concluded:
the effects that diet composition can have on metabolism and found that diets high in fat and low in carbohydrate do in fact lead to weight loss by increasing energy expenditure.
Remarkably, animals eating ketogenic diet lost a small amount of weight and achieved the same weight and body composition as animals that were calorie restricted to 66% of usual daily intake.
One problem with ketogenic LC-HF diets is that it is difficult to attribute observed effects (e.g., loss of body weight) to either the presence of ketone bodies or to the normally very low protein content of these diets.
The ideal ketogenic diet for research purposes would be a LC-HF diet that is ketogenic but ensures the sufficient supply of protein at the same time. However, until now, it is not clear whether the absence of dietary carbohydrates per se or the absence of carbohydrates in combination with a specific abundance of the two other macronutrients, fat and protein, is required to induce ketosis.
Peter’s interest is in fat storage based on insulin levels. The study compared ad libitum to calorie restricted eating in the mice. Peter concentrated on the ad libitum eating of the mice (not being all that interested in calorie restricted diets). Peter points out that it takes insulin to store fat in subcutaneous tissues but very little insulin to store fat in visceral fat. The study put it this way:
In all WAT depots, CR markedly upregulated the expression of proteins involved in FA biosynthesis in fed rats. In visceral WAT (rWAT and eWAT), hormone-sensitive lipase (lipolytic form) phosphorylation was increased by CR under fed conditions, and decreased by CR under fasted conditions. Conversely, in sWAT, hormone-sensitive lipase phosphorylation was increased by CR under fasted conditions. CR enhanced the effect of feeding on AKT activity in sWAT (indicative of a positive effect on insulin sensitivity) but not in rWAT or eWAT. These data suggest that CR improves lipid metabolism in an insulin signaling-dependent manner in sWAT only.
As Peter puts it:
This looks very much like one of the intrinsic differences between subcutaneous adipocytes and visceral adipocytes is that visceral adipocytes maintain insulin signalling at much lower levels of plasma insulin than do subcutaneous adipocytes. You have to store calories which arrive without insulin somewhere. Looks like this is the place!
Eleven people with type 2 diabetes (49.5 ± 2.5 years, BMI 33.6 ± 1.2 kg/m2, nine male and two female) were studied before and after 1, 4 and 8 weeks of a 2.5 MJ (600 kcal)/day diet.
Here are the results:
After 1 week of restricted energy intake, fasting plasma glucose normalised in the diabetic group (from 9.2 ± 0.4 to 5.9 ± 0.4 mmol/l; p = 0.003).
Insulin suppression of hepatic glucose output improved from 43 ± 4% to 74 ± 5% (p = 0.003 vs baseline; controls 68 ± 5%).
Hepatic triacylglycerol content fell from 12.8 ± 2.4% in the diabetic group to 2.9 ± 0.2% by week 8 (p = 0.003).
The first-phase insulin response increased during the study period (0.19 ± 0.02 to 0.46 ± 0.07 nmol min−1 m−2; p < 0.001) and approached control values (0.62 ± 0.15 nmol min−1 m−2; p = 0.42).
Maximal insulin response became supranormal at 8 weeks (1.37 ± 0.27 vs controls 1.15 ± 0.18 nmol min−1 m−2).
Pancreatic triacylglycerol decreased from 8.0 ± 1.6% to 6.2 ± 1.1% (p = 0.03).
Other interesting factoids from the study. In Type 2 diabetics:
Beta cell function declines linearly with time, and after 10 years more than 50% of individuals require insulin therapy.
Here’s the data from the study.
101.5 ± 3.4
103.7 ± 4.5
99.7 ± 4.5*
94.1 ± 4.3*
88.4 ± 4.3*†
33.4 ± 0.9
33.6 ± 1.2
32.3 ± 1.2*
30.5 ± 1.2*
28.7 ± 1.3*†
Fat mass (kg)
36.2 ± 2.7
39.0 ± 3.5
36.6 ± 3.6*
31.7 ± 3.7*
26.3 ± 4.0*
64.7 ± 3.8
64.7 ± 3.0
63.2 ± 3.1
62.4 ± 3.0*
62.1 ± 3.0*
Waist circumference (cm)
105.0 ± 1.5
107.4 ± 2.2
104.4 ± 2.2*
99.7 ± 2.4*
94.2 ± 2.5*†
Hip circumference (cm)
109.8 ± 2.4
109.5 ± 2.9
108.3 ± 2.7*
105.0 ± 2.6*
99.5 ± 2.6*†
0.96 ± 0.02
0.98 ± 0.02
0.97 ± 0.02
0.95 ± 0.01
0.95 ± 0.01
It is remarkable that the people lost mostly fat. The Fat Free Mass loss was only 2.6kg (about 6 lbs). The fat loss was 10 kg (about 22 lbs). That’s a pretty decent drop.
This was neither a Low Carb nor Low Fat diet. It was a restricted calorie diet (600 calories a day). The macros were 46.4% carbohydrate, 32.5% protein and 20.1% fat; vitamins, minerals and trace elements; 2.1 MJ/day [510 kcal/day]; Optifast; Nestlé Nutrition, Croydon, UK. This was supplemented with three portions of non-starchy vegetables such that total energy intake was about 2.5 MJ (600 kcal)/day.
It is remarkable how much fat was lost from the liver in just the first week.
Hepatic triacylglycerol content decreased by 30 ± 5% during week 1 of intervention (p < 0.001), becoming similar to control values (p = 0.75). It continued to decline throughout the intervention period to reach the normal range for non-obese individuals  (2.9 ± 0.2%; p = 0.003; Fig. 1), i.e. a total reduction of 70 ± 5%.
Most interestingly, the study after the study noted:
Following the intervention, participants gained 3.1±1.0 kg body weight over 12 weeks, but their HbA1c remained steady while the fat content of both pancreas and liver did not increase.
The conclusion matches my own hypothesis:
The data are consistent with the hypothesis that the abnormalities of insulin secretion and insulin resistance that underlie type 2 diabetes have a single, common aetiology, i.e. excess lipid accumulation in the liver and pancreas.
…a high-fructose diet in rhesus monkeys produces insulin resistance and many features of the metabolic syndrome, including central obesity, dyslipidemia, and infl ammation within a short period of time; moreover, a subset of monkeys developed type 2 diabetes
A Rhesus monkey used in the study is closer genetically to a human than the typical mouse study.
Numerous animal studies, mostly conducted in rodents, have shown that diets high in fructose produce metabolic perturbations associated with the metabolic syndrome and T2DM.
However, important metabolic differences exist between rodents and primates, particularly with respect to lipoprotein metabolism,the major site of lipogenesis (liver vs. adipose), and the physiology of thermogenesis.
Therefore, the results of metabolic studies performed in primates are substantively more applicable to human physiology and medicine than those from rodent studies, underscoring the importance of developing standardized nonhuman primate models of insulin resistance for the study of metabolic syndrome and T2DM.
Kimber Stanhope was one of the authors of this study. See her other study on Fructose (Not All [Sugar] Is Bad).
That’s a challenging question since by most tests I am not a diabetic. I no longer take diabetic meds and have good control of my blood sugars. The Type 2 Diabetes ADA Diagnosis Criteria are any of the following:
A hemoglobin A1c (HbA1c) level of 6.5% or higher; the test should be performed in a laboratory using a method that is certified by the National Glycohemoglobin Standardization Program (NGSP) and standardized or traceable to the Diabetes Control and Complications Trial (DCCT) reference assay, or
A fasting plasma glucose (FPG) level of 126 mg/dL (7 mmol/L) or higher; fasting is defined as no caloric intake for at least 8 hours, or
A 2-hour plasma glucose level of 200 mg/dL (11.1 mmol/L) or higher during a 75-g oral glucose tolerance test (OGTT), or
A random plasma glucose of 200 mg/dL (11.1 mmol/L) or higher in a patient with classic symptoms of hyperglycemia (ie, polyuria, polydipsia, polyphagia, weight loss) or hyperglycemic crisis
I have changed the bullet-ted list to a numbered list for convenience. I am on no diabetes medications to mask the results here:
My last HbA1C was 5.2 so I pass this test.
My fasting plasma glucose is less than 100 typically so I pass this test.
I have not had an OGTT (more on this to follow).
I have none of the symptoms of hyperglycemia at all and I have had no blood sugar measurements of 200 or higher (or anywhere near that level) since I started Low Carb.
Oral Glucose Tolerance Test (OGTT)
I don’t know if I would pass an OGTT or not. I assume I would fail such at test in spite of losing 120 lbs, etc. The reason I assume I would fail is that I think part of being on a Low Carb ketogenic diet is that my body has developed peripheral insulin resistance.
After 8 wk of consuming chow or KD, caloric intake after peripheral or central insulin and insulin and glucose levels after a glucose challenge were assessed. In a separate group of rats, glucose and insulin responses to either a low- or high-carbohydrate test meal were measured. Finally, rats maintained on KD were switched back to a chow diet, and insulin sensitivity and glucose tolerance were evaluated to determine whether the effects of KD were reversible.
That answers the test that I would want to do to determine if I would pass an OGTT. What happened to the mice?
Maintenance on KD resulted in decreased sensitivity to peripheral insulin and impaired glucose tolerance.
So after 8 weeks of not eating carbohydrates the mice had trouble eating carbohydrates. Not much of a surprise there. It would take a deeper dive to see how much worse their PIR and IGT became.
Furthermore, consumption of a high-carbohydrate meal in rats that habitually consumed KD induced significantly greater insulin and glucose levels for an extended period of time, as compared with chow-fed controls.
So the mice over-reacted to carbohydrate meals by producing more glucose and insulin.
Responsivity to central insulin was heightened in KD rats and associated with increased expression levels of insulin receptor mRNA.
Not sure how to understand that if the mice were more insulin resistant. But was this effect a permanent change or was it temporary and a side effect of the diet itself?
Finally, returning to a chow diet rapidly reversed the effects of KD on insulin sensitivity and glucose tolerance. These data suggest that maintenance on KD negatively affects glucose homeostasis, an effect that is rapidly reversed upon cessation of the diet.
Although 8 weeks isn’t that long to a human it’s a long time to a mouse. I don’t know the scaling factor but it’s reasonable to assume it is years rather than the two months of the study.
So, if someone is concerned about whether or not they would pass an OGTT it seems like they probably could stop the ketogenic diet for some time (weeks maybe?) and then take the test. Most of us who do LC / Keto won’t be trying it anytime soon.
The fact is your doctor is not going to order an OGTT for you if you don’t fail one or more of the other numbers. In fact, if you fail the fasting blood sugar test the doctor might order you an HbA1C test for confirmation. And then, depending on other factors, may just decide to keep an eye on it.
There’s a common definition of the word “obese”. We think of people who are really fat as being obese. I was one of them.
What is Obesity?
Obesity has a technical definition which is somewhat arbitrary. It is a function of weight and height and is known as BMI (Body Mass Index). The US government definition is (NCHS Data Brief ■ No. 288 ■ October 2017):
Obesity: BMI was calculated as weight in kilograms divided by height in meters squared, rounded to one decimal place.
Obesity in adults was defined as a BMI of greater than or equal to 30.
BMI Weaknesses as a Metric
BMI (and obesity) does not take into account body composition such as body fat or lean body mass. Two people can have the same BMI and be technically obese and one be solid muscle with little body fat and the other have significantly more body fat.
However, for the “average” person BMI is a decent measurement of fatness.
Obesity and Health
Generally, obesity and health are inversely related but there are people who are obese (by BMI) but are healthy. There are also people who are not obese but have poor health. This observation has led to the concept of personal fat threshold (PFT). This is described in (Taylor R, Holman RR. Normal weight individuals who develop type 2 diabetes: the personal fat threshold. Clin Sci (Lond). 2015 Apr;128(7):405-10) (PDF).
Personal Fat Threshold (PFT)
The Personal Fat Threshold concept is that there’s a level of fatness which the individual can tolerate before their health is impacted. This concept is tempting but has some problems.
PFT is not all that useful in the a-priori sense. There is no objective test to see if someone is at or near their PFT. Obesity isn’t useful as a metric. Neither is body fat level.
The only use of PFT is to support the medical advice to patients of weight loss as a tool for management of Type 2 diabetes. The PFT concept doesn’t actually contribute much since it has been believed (before the PFT concept was developed) that weight loss of about 15% resolves diabetes (Reversing Diabetes with Weight Loss: Stronger Evidence, Bigger Payoff).
Until there’s an a-priori means of measuring PFT the approach seems to be not all that useful. No medical doctor can tell you that you are 10 lbs away from your PFT. The point is completely hidden until it manifests. All it says that is if you are not technically considered to be obese and you are diabetic it is because you have gone over your personal fat threshold.
PFT – My Own Experience
There are three lines of reasoning from my own experience that call into question the PFT theory.
One was from my own experience with Insulin as a Type 2 Diabetic. I put on 40 lbs in a short time when I was put on Insulin. Conversely, when I got off Insulin my weight dropped quickly. Teenage females who are Type 1 diabetics and want to lose weight are well aware of this relationship. Weight increases followed Insulin increases (Skovsø S, Damgaard J, Fels JJ, Olsen GS, Wolf XA, Rolin B, Holst JJ. Effects of insulin therapy on weight gain and fat distribution in the HF/HS-STZ rat model of type 2 diabetes. Int J Obes (Lond). 2015 Oct;39(10):1531-8). not Insulin followed weight. Eventually, stasis is reached in weight and Insulin amount – at least in the short term.
Increasing dietary carbohydrates requires pumping more Insulin. When you stop eating dietary carbohydrates you don’t have to inject extra insulin for the meal.
The second reason was the increase in Insulin that is required over time to maintain blood sugar levels. I started at about 40g of Insulin and had good blood sugar controls. By four later my weight was stable but the amount of Insulin to keep blood sugar stable kept increasing to about 120 units. More particularly, the amount of insulin to cover carbohydrate loads increased. In my own case 1 unit of Insulin could cover 15 grams of carbs when I started Insulin and by four years later 1 unit wasn’t enough to cover 8 grams. All of this was at a stable weight (after the initial gain) and the same level of carbohydrates.
A third reason is my own weight history. I was at 285 lbs and non-diabetic for years. Then I mysteriously lost 50 lbs down to 235 lbs over the course of about six months. This is a common occurrence with Type 2 diabetics (Unexplained Weight Loss and Diabetes). After six months of this unexplained weight loss, I was then diagnosed with diabetes.
Perhaps this is the body pushing back from the PFT but it does call the concept into question – or at least indicate the real issue is much more complicated. After being put on Metformin my weight stabilized at around 10 lbs higher (although Metformin is said to lower weight). As my diabetes got worse my doctor tried different medications some of which added weight and some (like Byetta) caused small weight loss. Finally, the addition of Insulin added 40 lbs to my weight.
I did low carb while on Insulin but it only took my HbA1C down to 6.4. It wasn’t until I did low carb plus Intermittent Fasting that I was able to get off Insulin and my weight fell very quickly. My last HbA1C was 5.2 which is a normal non-diabetic number.
Carbohydrate Insulin Relationship
At the very least, if the PFT concept is salvageable, it needs to be modified for increasing Insulin Resistance levels. If the best treatment for diabetes is weight loss the best way for Type 2 Diabetics to lose weight is to reduce insulin levels. The best way to reduce insulin levels is to the insulin load of the diet. For a Type 2 Diabetic who is on Insulin this results in a loss of a lot of weight in a very short period of time.
The recommendation that losing 15% of body weight does not seem plausible to a diabetic like myself. I’ve lost more than 15% from my peak weight and not been able to control my diabetes. I lost weight with Low Carb by itself but not enough to get off Insulin. At it was more than 15% of weight loss. If I was told that losing 15% of my body weight would control my diabetes I would have told my doctor that I tried it and it didn’t work.
I lost much less than 15% of my weight in the beginning of Low Carb plus Intermittent Fasting and was able to get off Insulin completely. It was getting off Insulin which allowed me to lose weight. And it was reducing my body’s Insulin needs by the Low Carb diet and Intermittent Fasting which worked for me.
One explanation hypothesized for the greater weight loss on Low Carb diets is the Low Carb diet is said to have an inherent metabolic advantage. This metabolic advantage should manifest itself in a greater resting energy expenditure. The paper looked at two possible mechanisms – triglyceride cycling and glyceroneogenesis.
The critics of the Low Carb diet say that the advantage is that the comparisons aren’t done by holding protein constant. Overfeeding protein is not the same as overfeeding carbs or fat since protein stimulates 24 hour energy expenditure and fat doesn’t (Overfeeding Protein – Carnivore Diet).
Our results showed no significant weight loss, lipid, serum insulin, or glucose differences between the two diets.
The study was a decently formulated study but there were weaknesses:
Small study – only 4 subjects
No control group
Older obese females only
Very short duration (6 weeks)
Low fat didn’t get super-low (20% of calories from fat)
The Low Carb diet results in more weight loss but the study was too small to have statistical power
The good parts of the study were:
Matched total calories and protein – varying carbs and protein
Decent protein level (30% of calories)
Low carb was 5% of calories – good level
Randomized control trial
Cross-over design so the subjects ate both foods in random order
Starches and fruit were the carbohydrate choices (not jelly beans)
Deficit was relatively small (200 calories below REE which is a fairly large amount below TDEE depending on activity level)
At the end of the study they gave the participants the choice to continue on for a year. They were given the choice of the two diets and three of the four participants chose the Low Carb diet. However, the Low Carb participants raised their carbohydrate amount from 5% to 23% over the rest of the study so their weight loss partially reversed. There were several distinct advantages for the Low Carb diet.
Lipids were dramatically reduced on both diets, with a trend for greater triglyceride reduction on the VLC diet. Glucose levels were also reduced on both diets, with a trend for insulin reduction on the VLC diet
This fits my own experiences with protein and Low Carb. I’ve seen people stall for a long time and then break the stall by increasing their protein (and dropping their fat). My conclusion is that the ketogenic diet advantage does come from the higher protein intake of the diet. The diet often causes people to increase their consumption of meat which a high quality food.
This study showed that varying the protein content of several entrées consumed ad libitum did not differentially influence energy intake or affect ratings of satiety over a day. When the appearance, taste, fat content, and energy density were controlled, simply adding meat to lunch and dinner entrées to increase the protein content within commonly consumed amounts was not an effective strategy to reduce daily energy intake.